Optimal Life Insurance Decision in Mean-Variance DC Management with Mortality Improvements

Battle Count: 73

The paper is primarily focused on pension management and actuarial science rather than quantitative trading. However, it shares methodological foundations with portfolio optimization (mean-variance framework, martingale approach, stochastic control). The Vasicek interest rate model, market price of risk concepts, and efficient frontier analysis are directly relevant to fixed-income and multi-asset portfolio management. The treatment of mortality as a hedgeable risk and the insurance-as-asset-allocation framing are novel but less directly applicable to trading strategies. The analytical closed-form solutions and sensitivity analysis methodology could inform systematic portfolio construction.

The analytical framework involves solving a continuous-time stochastic control problem with multiple state variables (interest rate, contribution, mortality). The martingale approach requires constructing a mortality-adjusted pricing kernel and solving systems of ODEs for auxiliary functions. The closed-form solutions involve integrals over survival probabilities and exponential-affine functions of the short rate. Numerical implementation requires Euler-Maruyama discretization with Monte Carlo simulation. The Gompertz-Makeham mortality calibration adds another layer of complexity. However, the final formulas are explicit and computable given parameter inputs.

The paper provides detailed baseline parameters (Table 1), mortality model estimation (Table 2), and full analytical derivations in appendices. However, no code repository is provided. The Euler-Maruyama simulation scheme (400 time steps, 10,000 Monte Carlo paths) is described but not implemented publicly. Financial parameters are adopted from Munk and Sørensen (2010) rather than independently estimated. The SOA MIM-2021-v4 mortality data is publicly available but requires specific access.